| *** jcru has joined #openstack-trove | 00:19 | |
| *** sriram_tesora has quit IRC | 00:19 | |
| *** bhunter71 has quit IRC | 00:23 | |
| *** bhunter71 has joined #openstack-trove | 00:24 | |
| *** mattgriffin has quit IRC | 00:27 | |
| *** jcru has quit IRC | 00:28 | |
| *** pmalik has joined #openstack-trove | 00:41 | |
| *** bhunter71 has quit IRC | 01:01 | |
| *** topshare has joined #openstack-trove | 01:11 | |
| *** rwsu is now known as rwsu-afk | 01:17 | |
| *** bhunter71 has joined #openstack-trove | 01:18 | |
| *** nexusz99 has joined #openstack-trove | 01:19 | |
| *** IanGovett has quit IRC | 01:20 | |
| *** IanGovett has joined #openstack-trove | 01:21 | |
| *** pmalik has quit IRC | 01:31 | |
| *** rkourtz has quit IRC | 01:54 | |
| *** ashleighfarnham has quit IRC | 02:00 | |
| *** eghobo has joined #openstack-trove | 02:05 | |
| *** annashen has joined #openstack-trove | 02:07 | |
| *** erkules_ has joined #openstack-trove | 02:07 | |
| *** erkules has quit IRC | 02:10 | |
| *** annashen has quit IRC | 02:12 | |
| *** haomaiwang has joined #openstack-trove | 02:13 | |
| *** mattgriffin has joined #openstack-trove | 02:25 | |
| *** eghobo has quit IRC | 02:31 | |
| *** eghobo has joined #openstack-trove | 02:32 | |
| *** eghobo has quit IRC | 02:46 | |
| *** IanGovett1 has joined #openstack-trove | 02:59 | |
| *** IanGovett has quit IRC | 03:00 | |
| *** ashleighfarnham has joined #openstack-trove | 03:00 | |
| *** radez is now known as radez_g0n3 | 03:05 | |
| *** mattgriffin has quit IRC | 03:16 | |
| openstackgerrit | amrith proposed openstack/trove: Add short document on building guest images https://review.openstack.org/159627 | 03:16 |
|---|---|---|
| *** pmackinn has quit IRC | 03:20 | |
| *** IanGovett1 has quit IRC | 03:20 | |
| *** pmackinn has joined #openstack-trove | 03:21 | |
| *** pmackinn has joined #openstack-trove | 03:22 | |
| *** pmackinn has quit IRC | 03:22 | |
| *** IanGovett has joined #openstack-trove | 03:24 | |
| *** IanGovett1 has joined #openstack-trove | 03:26 | |
| *** IanGovett has quit IRC | 03:28 | |
| *** ashleighfarnham has quit IRC | 03:34 | |
| *** david-lyle is now known as david-lyle_afk | 03:44 | |
| openstackgerrit | amrith proposed openstack/trove-specs: Add a section for "Upgrade Implications" https://review.openstack.org/154578 | 03:52 |
| *** i-m-in has joined #openstack-trove | 03:53 | |
| *** IanGovett1 has quit IRC | 03:56 | |
| *** i-m-in has quit IRC | 03:57 | |
| *** amrith is now known as _amrith_ | 04:04 | |
| *** Longgeek has joined #openstack-trove | 04:08 | |
| *** _amrith_ is now known as amrith | 04:16 | |
| *** exploreshaifali has joined #openstack-trove | 04:50 | |
| *** ianbrown has joined #openstack-trove | 04:57 | |
| *** eghobo has joined #openstack-trove | 05:04 | |
| *** DCWilliams_VA has joined #openstack-trove | 05:09 | |
| *** DCWilliams_VA has quit IRC | 05:13 | |
| *** exploreshaifali has quit IRC | 05:14 | |
| *** ashleighfarnham has joined #openstack-trove | 05:20 | |
| *** ashleighfarnham has quit IRC | 05:24 | |
| *** eghobo has quit IRC | 05:38 | |
| *** eghobo has joined #openstack-trove | 05:38 | |
| *** exploreshaifali has joined #openstack-trove | 05:40 | |
| *** bhunter71 has quit IRC | 05:41 | |
| openstackgerrit | Sushil Kumar proposed openstack/trove: Add support for HP Vertica datastore in Trove https://review.openstack.org/156486 | 05:48 |
| *** ianbrown has quit IRC | 05:52 | |
| *** Riddhi has quit IRC | 06:06 | |
| *** coolsvap_ has quit IRC | 06:11 | |
| *** eghobo_ has joined #openstack-trove | 06:14 | |
| *** eghobo has quit IRC | 06:16 | |
| *** eghobo_ has quit IRC | 06:34 | |
| *** Longgeek has quit IRC | 06:39 | |
| openstackgerrit | Sushil Kumar proposed openstack/trove: Implement clustering for Vertica datastore https://review.openstack.org/159566 | 06:46 |
| *** nexusz99 has quit IRC | 06:47 | |
| openstackgerrit | Sushil Kumar proposed openstack/trove: Add support for HP Vertica datastore in Trove https://review.openstack.org/156486 | 06:49 |
| openstackgerrit | Sushil Kumar proposed openstack/trove: Implement clustering for Vertica datastore https://review.openstack.org/159566 | 06:49 |
| *** ianbrown has joined #openstack-trove | 07:11 | |
| *** Longgeek has joined #openstack-trove | 07:19 | |
| *** Longgeek has quit IRC | 07:43 | |
| *** ianbrown has quit IRC | 07:54 | |
| *** eghobo has joined #openstack-trove | 07:55 | |
| *** eghobo has quit IRC | 07:56 | |
| *** eghobo has joined #openstack-trove | 07:56 | |
| *** eghobo_ has joined #openstack-trove | 08:02 | |
| *** eghobo has quit IRC | 08:06 | |
| *** ashleighfarnham has joined #openstack-trove | 08:11 | |
| *** chlong has quit IRC | 08:11 | |
| *** ashleighfarnham has quit IRC | 08:16 | |
| *** eghobo_ has quit IRC | 08:49 | |
| *** erkules_ is now known as erkules | 09:04 | |
| *** jamie_h has joined #openstack-trove | 09:17 | |
| openstackgerrit | Merged openstack/trove-specs: Move the Trove Guest Agent to its own module https://review.openstack.org/136918 | 09:26 |
| *** Longgeek has joined #openstack-trove | 09:32 | |
| *** ianbrown has joined #openstack-trove | 09:51 | |
| *** ianbrown has quit IRC | 09:56 | |
| *** Longgeek has quit IRC | 10:06 | |
| *** Longgeek has joined #openstack-trove | 10:14 | |
| *** topshare has quit IRC | 10:14 | |
| *** tosky has joined #openstack-trove | 11:00 | |
| *** haomaiwang has quit IRC | 11:09 | |
| *** mattf is now known as _mattf | 11:13 | |
| *** chlong has joined #openstack-trove | 11:22 | |
| *** IanGovett has joined #openstack-trove | 11:51 | |
| *** ianbrown has joined #openstack-trove | 12:41 | |
| *** ianbrown has quit IRC | 12:46 | |
| *** DCWilliams_VA has joined #openstack-trove | 13:04 | |
| *** pboros has joined #openstack-trove | 13:17 | |
| *** amrith is now known as _amrith_ | 13:25 | |
| *** _mattf is now known as mattf | 13:28 | |
| *** pmackinn has joined #openstack-trove | 13:33 | |
| *** pmackinn has joined #openstack-trove | 13:35 | |
| *** ashleighfarnham has joined #openstack-trove | 13:39 | |
| *** ashleighfarnham has quit IRC | 13:44 | |
| *** radez_g0n3 is now known as radez | 13:46 | |
| *** exploreshaifali has quit IRC | 13:55 | |
| *** miqui has joined #openstack-trove | 13:56 | |
| *** DCWilliams_VA has quit IRC | 13:57 | |
| *** IanGovett has quit IRC | 14:01 | |
| *** IanGovett has joined #openstack-trove | 14:03 | |
| *** Longgeek has quit IRC | 14:05 | |
| *** fifieldt_ has quit IRC | 14:05 | |
| *** Longgeek has joined #openstack-trove | 14:20 | |
| *** _amrith_ is now known as amrith | 14:23 | |
| *** mattf is now known as _mattf | 14:30 | |
| *** chlong has quit IRC | 14:34 | |
| *** bhunter71 has joined #openstack-trove | 14:38 | |
| *** exploreshaifali has joined #openstack-trove | 14:56 | |
| *** sgotliv has quit IRC | 15:00 | |
| *** sgotliv has joined #openstack-trove | 15:00 | |
| amrith | vkmc, yt? | 15:08 |
| vkmc | amrith, here | 15:09 |
| amrith | Hi Victoria. If https://review.openstack.org/#/c/151780/7 is OK in principle and the only changes proposed are small, could we have this approved today before FFE. | 15:09 |
| amrith | I see your comments and it apperas that you agree in principle with all of what is written and the changes proposed are in the typo and rewrite for simplicity in reading realm | 15:10 |
| amrith | so, if that is the case, would you be OK with approving this and having those changes done as part of the submission? | 15:10 |
| vkmc | amrith, sure thing, I don't want to block it because of some minor comments | 15:10 |
| vkmc | let's move it forward | 15:11 |
| amrith | thanks vkmc, it is waiting for your +1 in that case | 15:11 |
| vkmc | :) thanks amrith | 15:12 |
| amrith | ie your WF +1 | 15:12 |
| *** radez is now known as radez_g0n3 | 15:12 | |
| amrith | vkmc, next question ... | 15:14 |
| amrith | re: https://review.openstack.org/#/c/154578/ | 15:14 |
| vkmc | amrith, done | 15:14 |
| vkmc | amrith, same reasoning applies | 15:14 |
| amrith | the reason I entered that line is that I want to make it explicit as a reminder to those who submit a change that involves an impact to upgrade, that documentation is in fact required. | 15:14 |
| vkmc | oh I see | 15:15 |
| amrith | the process for this isn't entirely clear | 15:15 |
| amrith | so there is a process that says "add DocImpact" to the chkin | 15:15 |
| amrith | but that doesn't always get documentation written. | 15:15 |
| amrith | hence I'm trying to make it explicit that there should be a bug in which the person submitting code also indicates what the upgrade documentation should look like. | 15:15 |
| amrith | or describe the change | 15:16 |
| amrith | in such a way that the technical writers have the information they need to write the documentation. | 15:16 |
| amrith | this has been called out a couple of times in reviews | 15:16 |
| amrith | but never 'formalized'. this was, maybe, my way of trying to formalize that. | 15:16 |
| amrith | in this way, when someone reviews the BP, one will see that there is an upgrade implication and therefore one must do something more. | 15:17 |
| amrith | that's the thought process. | 15:17 |
| amrith | not sure that it is right but that's why I wrote it that way. | 15:17 |
| openstackgerrit | Merged openstack/trove-specs: Trove Replication V2 https://review.openstack.org/151780 | 15:17 |
| vkmc | it makes sense yes | 15:17 |
| vkmc | I didn't think about it | 15:18 |
| vkmc | but I have seen many patches with the DocImpact flag that didn't get their corresponding docs | 15:18 |
| amrith | vkmc, to be clear, I'm not sure I'm right about this. | 15:18 |
| vkmc | hmm https://wiki.openstack.org/wiki/Documentation/DocImpact | 15:20 |
| vkmc | so apparently DocImpact should submit a bug to Documentation with the commit message of the change | 15:21 |
| amrith | the issue is that i've not seen any commit message that includes enough information for a tech writer. | 15:21 |
| vkmc | yeah | 15:22 |
| vkmc | perhaps we could stick to the usage of the docimpact flag, and ask the contributors to keep track of the opened bugs their change generates and add more information if needed | 15:23 |
| vkmc | maybe in that section you added we could mention that its expected the usage of the docimpact flag | 15:24 |
| amrith | sounds good, thanks | 15:26 |
| amrith | also, when you get a chance please take a look at https://review.openstack.org/#/c/161665/ | 15:26 |
| amrith | should be relatively quick | 15:27 |
| amrith | thx! | 15:27 |
| vkmc | cool :) | 15:28 |
| vkmc | sure! | 15:28 |
| *** ashleighfarnham has joined #openstack-trove | 15:29 | |
| *** dhellmann has joined #openstack-trove | 15:29 | |
| *** ashleighfarnham has quit IRC | 15:33 | |
| *** radez_g0n3 is now known as radez | 15:34 | |
| *** sriram_tesora has joined #openstack-trove | 15:38 | |
| *** vkmc is now known as cmkv | 15:41 | |
| *** pmalik has joined #openstack-trove | 15:42 | |
| amrith | thx | 15:46 |
| *** Riddhi has joined #openstack-trove | 15:47 | |
| openstackgerrit | amrith proposed openstack/trove-specs: Add a section for "Upgrade Implications" https://review.openstack.org/154578 | 16:10 |
| *** mattgriffin has joined #openstack-trove | 16:10 | |
| *** ianbrown has joined #openstack-trove | 16:18 | |
| *** ianbrown has quit IRC | 16:22 | |
| *** chlong has joined #openstack-trove | 16:33 | |
| *** i-m-in has joined #openstack-trove | 16:37 | |
| *** bhunter has quit IRC | 16:44 | |
| *** bhunter has joined #openstack-trove | 16:44 | |
| *** i-m-in has quit IRC | 16:47 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/trove: Updated from global requirements https://review.openstack.org/162227 | 16:54 |
| *** annashen has joined #openstack-trove | 17:22 | |
| *** eghobo has joined #openstack-trove | 17:27 | |
| *** eghobo has quit IRC | 17:27 | |
| *** eghobo has joined #openstack-trove | 17:28 | |
| *** rwsu-afk is now known as rwsu | 17:33 | |
| *** eghobo has quit IRC | 17:35 | |
| *** eghobo has joined #openstack-trove | 17:37 | |
| *** tosky has quit IRC | 17:43 | |
| *** david-lyle_afk is now known as david-lyle | 17:43 | |
| *** pmackinn has quit IRC | 17:44 | |
| *** ashleighfarnham has joined #openstack-trove | 17:46 | |
| *** eghobo has quit IRC | 18:00 | |
| *** sbfox has joined #openstack-trove | 18:02 | |
| *** eghobo has joined #openstack-trove | 18:06 | |
| *** annashen has quit IRC | 18:16 | |
| *** kfox1111 has joined #openstack-trove | 18:18 | |
| kfox1111 | is the guest_id the instance id or the trove id? | 18:18 |
| *** morganfainberg is now known as needscoffeebadly | 18:19 | |
| edmondk | the guest_id is the trove instance id | 18:19 |
| *** pmackinn has joined #openstack-trove | 18:20 | |
| edmondk | when doing a trove create this is file injected into nova through a nova boot and a file is created /etc/guest_info on the nova instance for the trove guest agent service to use | 18:21 |
| kfox1111 | hmm... ok. If I'm using a cloudinit template, how do I get it to populate /etc/guest_info? I tried pulling the instance_id from the metadata server, but its the wrong instance id apparaently. | 18:21 |
| kfox1111 | is the cloudinit script templated like the heat template? | 18:22 |
| kfox1111 | must I use clouddrive for that? | 18:22 |
| *** needscoffeebadly is now known as CaptainMorgan | 18:22 | |
| kfox1111 | should I just use the heat engine? Last I tried, it seemed little unstable. | 18:23 |
| edmondk | have you changed upstream trove in your repo to not file inject the guest_info? | 18:26 |
| edmondk | when you do a trove create etc/guest_info should already be there for you | 18:26 |
| *** sbfox has quit IRC | 18:26 | |
| *** coolsvap has joined #openstack-trove | 18:29 | |
| openstackgerrit | Duk Loi proposed openstack/trove: Inject guest conf files to configurable location https://review.openstack.org/100381 | 18:30 |
| *** kieleth has quit IRC | 18:32 | |
| kfox1111 | using rdo juno's trove. | 18:33 |
| kfox1111 | no etc/guest_info was created. | 18:33 |
| edmondk | are you getting active trove instances ? | 18:33 |
| kfox1111 | they start, and get stuck in building. | 18:34 |
| edmondk | etc/guest_info has been in for a long time from my understanding | 18:34 |
| edmondk | you could be having a problem with nova file injection not working | 18:34 |
| edmondk | can you test a nova boot with file injection | 18:34 |
| edmondk | without using trove | 18:34 |
| kfox1111 | must you use injection? | 18:35 |
| edmondk | currently yes | 18:35 |
| kfox1111 | with the heat template, it generates it trhough the template it lookcs like. | 18:35 |
| kfox1111 | ok. I'll double check that its enabled in trove too. | 18:35 |
| edmondk | I have actually been wanting to use nova metadata instead and convert everything to that instead of file injection | 18:35 |
| edmondk | have to put out a blueprint for it | 18:36 |
| edmondk | I am using it for the hot pools feature I did a demo for during mid cycle | 18:36 |
| edmondk | using metadata instead of file injection | 18:36 |
| kfox1111 | if you could template cloudinit the same way you can template the heat template, you could inject it that way too. | 18:36 |
| edmondk | can you cat out /var/logs/trove/trove-guestinfo.log | 18:36 |
| edmondk | to see what your error is and why trove is hanging on build | 18:37 |
| kfox1111 | it was failing due to nonexistant /etc/guest_info. | 18:37 |
| kfox1111 | let me see if I can get it enabled. | 18:37 |
| edmondk | yeah you can do it in cloudinit as well with a template either way | 18:37 |
| edmondk | yeah check nova file injection | 18:37 |
| edmondk | the nice thing about metadata is that you could do a nova show, and you would see the trove instance id in the metadata | 18:37 |
| edmondk | vs using cloudinit and heat | 18:37 |
| kfox1111 | true. | 18:38 |
| edmondk | which would still be a file on the instance where you have to ssh into the machine to find out | 18:38 |
| *** eghobo_ has joined #openstack-trove | 18:39 | |
| *** eghobo__ has joined #openstack-trove | 18:40 | |
| edmondk | This is an example when doing trove create of the file that is being injected: {'files': {'/etc/guest_info': u'[DEFAULT]\nguest_id=776b7a11-a509-4cac-9b9c-668f094acee6\ndatastore_manager=mysql\ntenant_id=9ed1af3e4599409294e296e7de24014b\n'} | 18:41 |
| edmondk | into nova create | 18:41 |
| kfox1111 | use_nova_server_config_drive = False. | 18:41 |
| kfox1111 | let me try again. :) | 18:41 |
| *** eghobo has quit IRC | 18:42 | |
| kfox1111 | arg. I'm up to 5 instances now stuck in building and not deletable. | 18:43 |
| kfox1111 | whats the procedure for making them deletable? | 18:43 |
| *** eghobo_ has quit IRC | 18:43 | |
| edmondk | 1. They will timeout :) and you can delete. 2. You can log into mysql->use trove; -> TRUNCATE table instances; , 3. You can increase your quota in trove | 18:45 |
| edmondk | if you manually delete the trove instances you need to make sure to delete the cinder volumes as well | 18:45 |
| edmondk | and nova instances | 18:45 |
| kfox1111 | 1, they don't. its been several days on some of them. | 18:47 |
| edmondk | the timeout is a value set from what I understand | 18:47 |
| kfox1111 | ouch. | 18:47 |
| edmondk | i would log into mysql | 18:47 |
| edmondk | and delete the rows | 18:47 |
| kfox1111 | could be. | 18:48 |
| kfox1111 | ok. :/ | 18:48 |
| edmondk | then cinder list delete the ones that available | 18:48 |
| kfox1111 | with the cloud init variant of things, is there a way to preload a security group rule in? (ssh access?) | 18:48 |
| kfox1111 | k. | 18:49 |
| edmondk | yeah you can add an ssh key in cloud init | 18:49 |
| edmondk | and have it write to authorized_keys | 18:49 |
| edmondk | you can put anything in there | 18:49 |
| kfox1111 | no, by default ssh is blocked to the vm's. every test vm I lauch, I have to go find the security group, poke ssh in so Ican debug it. | 18:49 |
| kfox1111 | its only allowing port mysql by default. | 18:50 |
| edmondk | on devstack you should be able to ssh by default | 18:50 |
| edmondk | port 22 should be open | 18:50 |
| edmondk | did you use redstack install? | 18:50 |
| edmondk | and kickstart mysql? | 18:50 |
| kfox1111 | ok, I'm seeing /dev/cdrom now. stsuff in it. don't have a /etc/guest_info out of the box. still looking.... | 18:51 |
| kfox1111 | packstack/rdo juno. | 18:51 |
| edmondk | is this through doing nova boot or trove create | 18:52 |
| edmondk | trove create injects it | 18:52 |
| edmondk | nova boot with your mysql image wont have guest_inf | 18:52 |
| edmondk | uess you create it through trove | 18:52 |
| kfox1111 | /tmp/mnt/openstack/content/0000 looks like its the right file. | 18:52 |
| kfox1111 | nope. using the trove ui in the dashboard. | 18:52 |
| kfox1111 | its like /etc/trove/cloudinit/mysql.cloudinit is overriding all the rest of the trove userdata logic. | 18:53 |
| edmondk | k | 18:53 |
| edmondk | is this the upstream mysql.cloudinit | 18:53 |
| edmondk | or do you have a custom one | 18:54 |
| kfox1111 | not sure. one of my co-workers got it. is there one upstream I can look at? hmm... | 18:54 |
| kfox1111 | https://github.com/openstack/trove/tree/master/etc/trove/cloudinit is empty. | 18:54 |
| edmondk | give me 5 minutes finishing up a meeting | 18:54 |
| edmondk | and will see if I can find it | 18:55 |
| kfox1111 | k. thx. | 18:56 |
| kfox1111 | well, manuallly copying the content/0000 file to /etc/guest_info and restarting trove-guest agent made the instance drop out of building and into error state. so that's progress | 18:58 |
| kfox1111 | I see nothing but "ERROR" state htough when showing the instance... | 18:58 |
| kfox1111 | Failed to create instance e0ea25aa-a6ff-4a33-93d7-c8bddf59c5e5. Timeout waiting for instance to become active. No usage create-event was sent. | 19:00 |
| kfox1111 | may be releated... | 19:00 |
| edmondk | interesting | 19:00 |
| edmondk | are you running on default devstack? | 19:02 |
| kfox1111 | ok. I just put some logic in the cloud init file to moun the config drive, and put guest_info in place. We'll see how a new launch goes. | 19:02 |
| edmondk | how did you set up your environment | 19:02 |
| kfox1111 | rdo juno + packstack. | 19:02 |
| edmondk | ok | 19:03 |
| kfox1111 | I'm guessing that trove normally generates some user-data script stuff, but if you use cloud-init scripts, it overrides the defaults entirely. | 19:03 |
| edmondk | the cloud-init is your userdata | 19:03 |
| edmondk | there is no userdata by default | 19:04 |
| edmondk | on devstack and trove | 19:04 |
| edmondk | it will be None | 19:04 |
| kfox1111 | that or cloud-init for some reason is ignoring user-data on config drive. | 19:04 |
| kfox1111 | then how does /etc/guest_info get populated? | 19:04 |
| kfox1111 | something's got to populate it. | 19:04 |
| edmondk | file injected | 19:04 |
| edmondk | {'files': {'/etc/guest_info': u'[DEFAULT]\nguest_id=776b7a11-a509-4cac-9b9c-668f094acee6\ndatastore_manager=mysql\ntenant_id=9ed1af3e4599409294e296e7de24014b\n'}, 'userdata': None, | 19:04 |
| edmondk | when I put a breakpoint on trove create | 19:04 |
| kfox1111 | thats data. what tells it to put the content in /etc/guest_info? | 19:04 |
| edmondk | right before nova servers create | 19:04 |
| edmondk | this is what is passed in | 19:04 |
| edmondk | that files keyname | 19:05 |
| edmondk | tells nova to file inject it | 19:05 |
| kfox1111 | seems like there is something inside the vm that reads that file, and puts it into the right place. I'm guessing thats cloud-init? My cloud-init does not seem to be touching the config drive at all. | 19:05 |
| edmondk | cloud init doesn't do anything with etc/guest_info | 19:06 |
| kfox1111 | the config drive is just a iso image stuck in /dev/cdrom. | 19:06 |
| edmondk | nova.servers.create in taskmanager of trove does that file injection | 19:06 |
| edmondk | using the above input | 19:06 |
| kfox1111 | yeah. I see the file, in the /dev/cdrom iso. | 19:06 |
| *** jamie_h has quit IRC | 19:07 | |
| kfox1111 | no where else though. | 19:07 |
| *** cmkv is now known as vkmc | 19:09 | |
| vkmc | kfox1111, hi there! you are using the script I shared with you? | 19:10 |
| kfox1111 | not yet. | 19:10 |
| kfox1111 | am really close with ubuntu at the moment. Just want to get something working to validate I have all the right config bits. then will try rhel images. | 19:11 |
| vkmc | kfox1111, sure | 19:11 |
| *** annashen has joined #openstack-trove | 19:11 | |
| vkmc | if it serves for anything... trove looks for guest_info and trove-guestagent.conf under /etc/ | 19:12 |
| vkmc | try putting there | 19:12 |
| vkmc | putting them* | 19:12 |
| vkmc | in the host | 19:12 |
| kfox1111 | yeah. I'm trying to get guest_info from trove though. I had config drive turned off in the config apparently. | 19:13 |
| kfox1111 | I turned it on, and am seeing the data for guest_info in the filesystem in /dev/cdrom, | 19:13 |
| vkmc | ok | 19:13 |
| kfox1111 | but nothing is putting it into /etc/guest_info | 19:13 |
| kfox1111 | I'm manually tewaking my config-init script to mount/copy it into place. | 19:13 |
| kfox1111 | didn't qui9te work, but if I ssh in, put the file in place, and start trove guest agent, it starts now, look like its waiting for something to do. | 19:16 |
| kfox1111 | still in building right now. | 19:16 |
| edmondk | have you checked the cloud-init logs? | 19:18 |
| kfox1111 | looks like its close... 2015-03-06 19:23:36.380 4273 INFO trove.guestagent.datastore.mysql.service [-] Service Status is RUNNING. | 19:18 |
| kfox1111 | darn. gota go. I'll be back in a bit. | 19:18 |
| edmondk | k | 19:18 |
| kfox1111 | actualy, change of plans. I can stay. | 19:19 |
| kfox1111 | so... with the guest info copying, the guest agent seems pretty happy. the task manager doesn't seem to know that its up yet though. | 19:21 |
| *** coolsvap has quit IRC | 19:25 | |
| kfox1111 | is there a communication channel somewhere that could be misconfigured? | 19:26 |
| *** sbfox has joined #openstack-trove | 19:29 | |
| kfox1111 | what does the notification driver do? | 19:33 |
| *** eghobo__ has quit IRC | 19:35 | |
| kfox1111 | there's an option, notifier_queue_hostname in the guest agent. is that used to talk to trove-taskmanager? | 19:35 |
| kfox1111 | ah. notifications.info queue has no consumers. | 19:44 |
| kfox1111 | could that be it? does trove taskmanager need to be listening there? | 19:45 |
| *** CaptainMorgan is now known as morganfainberg | 19:50 | |
| kfox1111 | hmm... guestagent.c3feecb1-0b35-4d9c-8bc5-13f0c7f873f6 has one posted message. no consumers. | 19:50 |
| kfox1111 | it seems to contain if for the guestagent? | 19:52 |
| kfox1111 | I'm not sure if its the client or the server not speaking. :/ | 19:53 |
| kfox1111 | ideas? | 19:53 |
| *** ianbrown has joined #openstack-trove | 19:56 | |
| kfox1111 | is trove ok with having its own rabbit? | 19:58 |
| *** ianbrown has quit IRC | 20:00 | |
| *** amrith is now known as _amrith_ | 20:04 | |
| kfox1111 | how odd. I can see messages flow from the guest agent to conductor. | 20:05 |
| kfox1111 | things are staying in building state. | 20:05 |
| kfox1111 | guestagent has mysql launched, but isn't creating the db. | 20:05 |
| kfox1111 | notifications.info keeps growing very slowly. | 20:06 |
| *** Longgeek has quit IRC | 20:06 | |
| *** sriram_ has joined #openstack-trove | 20:07 | |
| *** sriram_tesora has quit IRC | 20:10 | |
| *** exploreshaifali has quit IRC | 20:13 | |
| openstackgerrit | Merged openstack/python-troveclient: Clean up remaining references to TROVE_URL https://review.openstack.org/161665 | 20:14 |
| *** radez is now known as radez_g0n3 | 20:32 | |
| *** exploreshaifali has joined #openstack-trove | 20:45 | |
| SlickNik | kfox1111: reading scrollback | 20:46 |
| SlickNik | kfox1111: So from what I read it looks like the guestagent isn't able to receive notifications from the taskmanager. | 20:47 |
| SlickNik | The purpose of getting /etc/guestinfo on the guest is to have a way of letting the guest agent know which queue it needs to listen on once it's started up (which is essentially guestagent.<trove-instance-id>) | 20:49 |
| SlickNik | The guest will not go from a BUILD to an active state until it receives the first "prepare" message from the taskmanager on that queue (in this case guestagent.c3feecb1-0b35-4d9c-8bc5-13f0c7f873f6) | 20:50 |
| kfox1111 | yeah. I have the guest_info file being populated now. it has a guestagent.uuid in rabbit. one entry. no listeners. The guestagent started mysql, is periodically pinging it,and calling out to conductor. | 20:55 |
| kfox1111 | otherwise, it seems stuck. | 20:55 |
| *** eghobo has joined #openstack-trove | 20:55 | |
| kfox1111 | ok. so, I need to figure out why the guest agent isnint picking up the message in guestagent.c3feecb1-0b35-4d9c-8bc5-13f0c7f873f6, but is somehow able to contact out to the conductor on the same rabbit. :/ | 20:56 |
| *** eghobo_ has joined #openstack-trove | 20:58 | |
| *** eghobo has quit IRC | 21:01 | |
| kfox1111 | does the guest agent stay connected to rabbit, or does it pop in and out? | 21:03 |
| kfox1111 | ok, so I restarted the guest agent, and am looking at the client that poped into rabbit. | 21:07 |
| kfox1111 | its listening on 3 queues: guestagent, guestagent.0.0.0.0 and guestagent_fanout_somenumber | 21:08 |
| kfox1111 | it has one more client, that isn't listening on anything. maybe just a sender? | 21:09 |
| kfox1111 | at any rate, its not listening on its queue. | 21:09 |
| *** pmackinn has quit IRC | 21:10 | |
| kfox1111 | ah.... I see a version missmatch. | 21:11 |
| kfox1111 | the guestagent is from icehouse. :/ | 21:11 |
| kfox1111 | the server is juno... | 21:11 |
| kfox1111 | let me see if I can fix that. | 21:11 |
| *** annashen has quit IRC | 21:14 | |
| *** annashen has joined #openstack-trove | 21:14 | |
| kfox1111 | so, is there trove guestagent juno package for ubuntu anywhere? | 21:20 |
| kfox1111 | :/ | 21:20 |
| kfox1111 | 14.10 only it seems. :/ | 21:22 |
| openstackgerrit | Duk Loi proposed openstack/trove: Inject guest conf files to configurable location https://review.openstack.org/100381 | 21:25 |
| kfox1111 | no, 14.10 seems to be unhappy too. | 21:34 |
| *** eghobo_ has quit IRC | 21:35 | |
| kfox1111 | hmm.. 14.10 doesn't seem to be listening on any queues. :/ | 21:36 |
| *** eghobo has joined #openstack-trove | 21:38 | |
| kfox1111 | so, yeah, its not listening on the queues it needs to. what other things might affect that? | 21:52 |
| kfox1111 | I see 2 clients belonging to this guest agent. one is writing something periodically. probably to conductor. the other seems idle. not listening on anything. | 21:55 |
| *** exploreshaifali has quit IRC | 21:56 | |
| kfox1111 | ah... maybe progress... it didn't seem to be honoriing /etc/guest_info. :/ | 22:04 |
| kfox1111 | I crudini merged it into guestagent.conf and it seems to be doing different things now. | 22:04 |
| *** mattgriffin has quit IRC | 22:07 | |
| *** radez_g0n3 is now known as radez | 22:08 | |
| *** pboros has quit IRC | 22:13 | |
| *** sbfox has quit IRC | 22:17 | |
| *** sbfox has joined #openstack-trove | 22:21 | |
| *** eghobo has quit IRC | 22:22 | |
| *** chlong has quit IRC | 22:24 | |
| *** exploreshaifali has joined #openstack-trove | 22:26 | |
| *** ianbrown has joined #openstack-trove | 22:27 | |
| *** radez is now known as radez_g0n3 | 22:31 | |
| *** ianbrown has quit IRC | 22:32 | |
| *** IanGovett has quit IRC | 22:36 | |
| *** sbfox has quit IRC | 22:37 | |
| *** sbfox has joined #openstack-trove | 22:44 | |
| *** kfox1111 has quit IRC | 22:56 | |
| *** sriram_ has quit IRC | 23:08 | |
| *** eghobo has joined #openstack-trove | 23:10 | |
| *** exploreshaifali has quit IRC | 23:12 | |
| *** pmalik has quit IRC | 23:16 | |
| *** pmalik has joined #openstack-trove | 23:19 | |
| *** pmalik has quit IRC | 23:20 | |
| *** annashen has quit IRC | 23:23 | |
| *** Riddhi has quit IRC | 23:28 | |
| *** amcrn has joined #openstack-trove | 23:48 | |
| *** annashen has joined #openstack-trove | 23:52 | |
| *** amcrn has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!